Provisioned throughput
Provisioned throughput is reserved capacity for a deployment, as opposed to pay-as-you-go Standard. When a model retires, Azure does not auto-upgrade provisioned deployments, and Bedrock blocks new Provisioned Throughput for a model in Legacy.
Last reviewed . Publisher: SunsetDiff. A source line names the day that page was read.
The rows below are copied from this site's model records. A shutdown cell is the date stored for that provider, not a date copied from another host.
Azure OpenAI
Azure lists Global Provisioned as reserved throughput with global routing, after Global Standard. Data Zone Provisioned keeps processing inside a geographic boundary. Regional Provisioned comes later.
Provisioned deployments are not auto-upgraded. Customers migrate them manually. The lifecycle page describes an in-place migration over a 20 to 30 minute window, or a side-by-side migration with a second deployment. Standard quota is handled on a Standard auto-upgrade. That sentence is about Standard, not provisioned quota.
Source: Foundry Models lifecycle , checked October 22, 2026.
Amazon Bedrock
On the current lifecycle page, you cannot create a new Provisioned Throughput for a model in the Legacy state. If you already customized the model before Legacy, you can keep using an existing Provisioned Throughput endpoint that was created before Legacy. You can also create a new custom model deployment for on-demand inference.
The dataset does not store whether your account uses provisioned capacity. The model row is the lifecycle state, not your deployment type.
Source: Bedrock model lifecycle , checked October 22, 2026.
OpenAI, Anthropic, and Gemini
The OpenAI, Anthropic, and Gemini deprecations pages read for this glossary do not define provisioned throughput. They retire a model id. Azure and Bedrock are the hosts that attach a capacity type to that id.
Sources: OpenAI deprecations , Anthropic model deprecations , Gemini API deprecations , checked October 22, 2026.
Rows from this dataset
The dataset has no deployment-type field. These are the model rows the provider pages use as lifecycle examples.
| Model | Provider | Stored status | Shutdown | Replacement |
|---|---|---|---|---|
gpt-4o (2024-05-13) gpt-4o-2024-05-13 | Azure OpenAI | deprecated | 2026-12-09 | gpt-5.6-sol ยท diff |